JetBrains Central: The Governance Layer That Shows Up When You're Juggling Claude Code, Codex, and Junie on the Same Team
JetBrains just announced Central, an open platform for unified AI agent management — and the timing reflects a specific operational crisis that's only visible once a team gets past the "our first agent works" milestone. The problem isn't agent capability. It's what happens when the same engineering team is running Claude Code, Codex, and Junie simultaneously on related codebases, with different permission models, different cost structures, and no shared visibility into what any of them are doing.
Zen van Riel's analysis, drawn from multi-team implementation experience, maps Central to three architectural layers. Governance and Control handles policy enforcement, identity management, cost attribution, and audit trails — the compliance infrastructure that large organizations require before they can formally adopt agents at all. Agent Execution Infrastructure provides cloud runtimes that handle sandboxing, resource allocation, and failure recovery, removing the per-team setup burden that makes multi-agent coordination expensive to operate. Agent Optimization addresses something subtler: when multiple agents operate on related codebases, they currently rebuild context independently, starting from scratch each time. Central's shared semantic context layer lets accumulated knowledge persist and transfer across agents and repositories, so the second agent to touch a service isn't working blind to what the first one already learned.
The market framing embedded in the article is striking. JetBrains' AI Pulse survey finds that 90% of developers already use AI at work, but only 22% use coding agents — with 66% of companies planning adoption in the next 12 months. That gap between current and planned adoption is not, the analysis argues, a capability problem. It's an operational complexity problem, and it's what Central is designed to close. A major IDE vendor entering the governance-first platform space rather than building another coding feature is a category signal worth tracking.
Central explicitly supports Claude Agent, Codex, Gemini CLI, and Junie side-by-side, positioning it as infrastructure that sits above any single vendor's agent rather than as a replacement for the agents themselves.